In the immortal words of Nathan Scott (nathans@xxxxxxx):
>
> On Sat, Jan 12, 2002 at 12:21:44PM -0500, Nathan J. Mehl wrote:
> > ...
> > At this point, I am stymied. xfs_repair will not progress beyond the
> > log-clearing step.
>
> Can you tell me the exact error message you see?
>
> Is it:
> "xlog_find_tail returned error X"?
Yup! The error was, and is:
Phase 1 - find and verify superblock...
sb root inode value 18446744073709551615
inconsistent with calculated value 13835051870529257600
resetting superblock root inode pointer to 18446744069414584448
sb realtime bitmap inode 18446744073709551615
inconsistent with calculated value 13835051870529257601
resetting superblock realtime bitmap ino pointer to 18446744069414584449
sb realtime summary inode 18446744073709551615
inconsistent with calculated value 13835051870529257602
resetting superblock realtime summary ino pointer to
18446744069414584450
Phase 2 - using internal log
- zero log...
XFS: Log inconsistent (didn't find previous header)
XFS: failed to find log head
fatal error -- xlog_find_tail returned error 5
> If so, I think this is a recently-introduced repair bug - I'll
> check in a fix shortly & you should be able to proceed on with
> xfs_repair.
Okay, I am a touch frightened to do just that. The output from
xfs_repair -n is, uh, copious. At ~600k, I won't spam the list with
it, but it can be perused here: http://blank.org/memory/xfs_repair.out.txt
I'm especially worried about:
entry "tmp" at block 0 offset 96 in directory inode 128 references
non-existent inode 20971648
would clear inode number in entry at offset 96...
entry "dev" at block 0 offset 112 in directory inode 128 references
non-existent inode 25165952
would clear inode number in entry at offset 112...
entry "bin" at block 0 offset 160 in directory inode 128 references
non-existent inode 67109191
would clear inode number in entry at offset 160...
entry "home" at block 0 offset 176 in directory inode 128 references
non-existent inode 83886263
would clear inode number in entry at offset 176...
entry "lib" at block 0 offset 192 in directory inode 128 references
non-existent inode 88080769
would clear inode number in entry at offset 192...
entry "mnt" at block 0 offset 208 in directory inode 128 references
non-existent inode 96469292
would clear inode number in entry at offset 208...
entry "opt" at block 0 offset 224 in directory inode 128 references
non-existent inode 109052059
would clear inode number in entry at offset 224...
entry "root" at block 0 offset 240 in directory inode 128 references
non-existent inode 113246387
would clear inode number in entry at offset 240...
entry "sbin" at block 0 offset 256 in directory inode 128 references
non-existent inode 117440698
would clear inode number in entry at offset 256...
entry "misc" at block 0 offset 272 in directory inode 128 references
non-existent inode 75508160
There are quite a lot of errors of that sort, but right there it looks
like my entire root directory is toast.
Is there anything that can be done manually to avoid this, or should I
just let xfs_repair do its job and salvage what I can from lost+found?
-n
------------------------------------------------------------<memory@xxxxxxxxx>
SENDING JUNK EMAIL TO MY ADDRESS CONSTITUTES YOUR LEGALLY-BINDING ACCEPTANCE
OF MY OFFER TO REMOVE BOTH OF YOUR NIPPLES WITH AN ORBITAL SANDER.
(--Andy Ihnatko)
<http://blank.org/memory/>----------------------------------------------------
|